Prognostic Factor for Subacute Encephalitis with First-Episode Psychosis

Encephalitis is elicited by aberrant inflammatory and immunological responses in the brain. Although symptomatic manifestations vary, they typically include seizures, memory deficit(s), and altered consciousness. Psychosis also occurs as an initial manifestation, especially among those with limbic encephalitis or autoimmune encephalitis. In clinical practice, clinicians often encounter the question of whether first-episode psychosis (FEP) originates from encephalitis itself or if encephalitis presenting with FEP develops concurrently. Previous studies have reported that such patients with FEP had a low prevalence of anti-neuronal antibodies [1], and FEP with serum N-methyl-D-aspartate (NMDA) receptor antibodies would respond poorly to antipsychotic medication [2]. Nevertheless, the prognosis of FEP among patients with overall encephalitis, including autoimmune encephalitis, remains uncertain.

Signs of aberrant inflammatory and immunological responses among patients with FEP may include behavioral problems, cognitive impairment, neurological soft signs, and autonomic abnormalities [3]. In such cases, experts in the field of autoimmune encephalitis have recommended lumbar puncture, electroencephalogram (EEG) examination, or administration of cerebrospinal fluid (CSF) neuronal autoantibodies as a subsequent procedure [4]. Although CSF examination is the most precise method for detecting inflammation in the central nervous system, even when the results of cerebral magnetic resonance imaging (MRI) or EEG are normal, these studies are rare in patients with psychosis [3]. Pleocytosis has been observed in <5% of patients in cohorts with undifferentiated psychosis [5, 6]; however, these cohorts included patients without encephalitis or CSF pleocytosis. If a patient with FEP exhibits an increase in white blood cell count in the CSF, its diagnosis will be confirmed as encephalitis.

There are few classifications of encephalitis in International Classification of Diseases-10 (G04). Other encephalitis (G04-8) includes anti-NMDAR encephalitis and autoimmune limbic encephalitis, in addition to five diseases. The remaining types of encephalitis are summarized as unclassifiable and unknown encephalitis (G04-9). This may be because encephalitis has a varied etiology, with many cases having unknown causes. At the stage of encephalitis showing FAP as an initial symptom, many facilities often do not provide test results indicating autoantibodies and causative viruses required for classification according to International Classification of Diseases. Thus, examining encephalitis as a whole is important and meaningful.

We examined a prognostic factor in patients with encephalitis who had both FEP and CSF pleocytosis. Verifying this may be helpful in clinical management, such as reconsideration of therapy or informed consent.

A total of 36 patients (mean age ±SD, 43.7 ± 19.3 years, range 16–76 years; 19 females), who presented with FEP from among a group of patients with subacute encephalitis undergoing treatment at Nara Medical University Hospital (Kashihara, Nara, Japan), were enrolled. Psychosis is defined as a significant impediment to reality testing as hallucinations, delusions, disassembled (disorganized) conversations, and disassembled (disorganized) behaviors [7]. Inclusion criteria were increased white blood cell count ≥5 in the CSF [8] and the presence of psychosis without neurological manifestations such as seizure or involuntary movement at the disease onset. Two patients did not exhibit pleocytosis ≥5 at the initial CSF examination; however, a follow-up CSF examination showed a >5 times increase in the white blood cell count. All patients were examined for viral antibodies using screening assays, including those for herpes simplex virus (HSV), and CSF culture testing. Polymerase chain reaction (PCR) confirmed HSV in 32 patients and varicella-zoster virus (VZV) in 6. Eleven patients were examined for cell surface antibodies. The etiology of encephalitis was confirmed in 14 patients (HSV [n = 3], VZV [n = 1], anti-NMDA receptor antibody [n = 9], and anti-leucine-rich glioma-inactivated 1 antibody [n = 1]). The dosage and duration of steroid treatment were left to the discretion of the treating physicians. All patients received acyclovir as the first treatment for suspected HSV encephalitis; however, therapy was discontinued when PCR results for HSV or VZV were negative. Exclusion criteria were as follows: detection of bacterial, tuberculosis, or fungal origin on CSF culture; detection of syphilis on serological testing; the presence of suspected brain tumor on computed tomography or initial and follow-up MRI as assessed by experimental radiologists; subsequent diagnosis of brain tumor; history of systemic immunological disease; or the presence of acute clinical episodes after vaccinations or clinical multiple relapsing episodes with abnormal white matter lesions in the central nervous system, suggesting acute disseminated encephalomyelitis or multiple sclerosis. In addition, the results of serological and radiological investigations ruled out the following diseases and conditions: cerebral thromboembolism and hemorrhage; cerebral venous sinus thrombosis; cerebral aneurysm; cerebral arteriovenous malformation; mitochondrial encephalopathy; diabetic encephalopathy; hepatic encephalopathy; systemic lupus erythematosus; antiphospholipid antibody syndrome; sarcoidosis; and Hashimoto encephalopathy.

A score of ≥3 and ≤2 on the modified Rankin scale [9] at 2 months after the onset were defined as poor and good outcomes, respectively. A total of 13 independent variables were evaluated and scored, where applicable, as follows: age; sex (female score = 1); detection of virus or autoantibodies etiology (absent = 0, present = 1); time from the neurological onset to initiation of acyclovir treatment; Glasgow Coma Scale score at initiation of acyclovir treatment; leukocyte count in the CSF (/mm³); protein level in CSF (mg/dL); detection of focal lesions on initial computed tomography (absent = 0, present = 1); focal lesions on cranial MRI (absent = 0, present = 1); detection of periodic lateralized epileptiform discharges (i.e., PLEDs) on EEG (absent = 0, present = 1); generalized seizure during the disease course (absent = 0, present = 1); use of mechanical ventilation at the acute stage (not administered = 0, administered = 1); and use of steroid treatment(s) at the acute stage (not administered = 0, administered = 1).

Variables demonstrating a statistically significant association with poor outcome(s) on univariate logistic regression analysis (p < 0.05) were entered into the multivariate logistic regression analysis using forced entry. Odds ratio (OR) and corresponding 95% confidence interval (CI) was calculated. Correlations of each variable were evaluated using Spearman’s rank correlation test. Differences in these 13 clinical characteristics between patients with poor outcomes and those with good outcomes were evaluated using the Mann-Whitney test or χ² test. Statistical analysis was performed using SPSS version 24 (IBM Corporation, Armonk, NY, USA).

The clinical characteristics of 36 patients with subacute encephalitis exhibiting FEP are summarized (Table 1). Detailed information regarding FEP is summarized in online supplementary Table 1 (for all online suppl. material, see www.karger.com/doi/10.1159/000524975). Significant variables on univariate logistic regression analysis included female sex (OR 5.571, 95% CI: 1.297–23.934; p = 0.021) and the use of mechanical ventilation during the acute stage (OR 7.286, 95% CI: 1.508–35.211; p = 0.013) (Table 2). On multivariate logistic regression analysis adjusting for the mechanical ventilation use and sex, the use of mechanical ventilation during the acute stage (OR 5.446, 95% CI: 1.044–28.615; p = 0.044) was significantly associated with poor outcomes. There were no significant interactions between these predictors. Furthermore, age – which has been reported to be a risk factor for poor prognosis in patients with subacute encephalitis [10] – was entered into the multivariate logistic regression analysis. Similarly, the use of mechanical ventilation during the acute stage demonstrated an OR of 5.840 (95% CI: 1.069–31.912; p = 0.042). The proportion of use of mechanical ventilation at the acute stage and female sex was higher among patients with poor outcomes than in those with good outcomes (p = 0.014 and p = 0.023, respectively). Other variables did not differ between the groups. The characteristics of patients who experienced poor outcomes are summarized in Table 3.

Results of the present study demonstrated that the use of mechanical ventilation was significantly associated with poor prognosis among patients with subacute encephalitis with FEP. A previous retrospective study, involving 195 patients with subacute encephalitis that examined outcome at discharge and 1 year after the disease onset, reported that the use of mechanical ventilation was an independent risk factor for poor prognosis [11]. This cohort included 139 patients with various etiologies including viral encephalitis (48%) or autoimmune encephalitis (22.2%), similar to our study. The proportion of patients with FEP was not reported; however, confusion at clinical presentation was high, with an estimate of approximately 70% [11]. Disease etiology in 22 (61%) patients in our study was unknown. However, undetermined cause remains as high as > 50% in acute encephalitis [12]; therefore, reliable risk factors for poor prognosis in acute etiology-undetermined encephalitis may be useful in clinical practice.

Female sex was found to be a risk factor for a poor outcome among those with encephalitis with FEP on univariate analysis. Most such studies have failed to find a difference in sex as a prognostic factor, although one study investigating anti-leucine-rich glioma-inactivated 1 encephalitis reported that female sex was significantly associated with poor prognosis at both complete follow-up and 1 year after the disease onset [13]. The authors reported that a stronger immune response among females compared with males could explain the association between female sex and poor prognosis.

The cause of encephalitis was not significantly associated with a poor outcome in our study. Previous studies involving patients with FEP are limited. One study (n = 180 patients) with a CSF pleocytosis rate of 3.4% reported that the positive rate of serum neuronal cell surface antibodies was not significantly different between patients with FEP and controls, with rates of 9% and 4%, respectively [5]. The positive patients did not develop encephalitis or neurological symptoms. Among 81 patients with schizophrenia spectrum disorders and a 4.8% prevalence of CSF pleocytosis, three (3.7%) patients had serum low-titer neuronal antibodies against different surface antigens and did not have an abnormal neurological examination suggestive of encephalitis [6]. Twenty of 228 (9%) patients with FEP with different positive serum surface antibodies did not experience a poorer outcome compared with seronegative patients who did not develop encephalitis during a 6-month follow-up period [14]. Limbic encephalitis involves the amygdala in the limbic system as the core region of neuropsychiatric manifestations. Viral etiology was not significantly associated with poor outcomes in 31 acyclovir-treated patients with subacute limbic encephalitis [15]. These observations are likely to support the absence of a relationship between prognosis in overall encephalitis and etiology. However, it is important to recognize the prognosis associated with each etiology when initially treating a patient with both CSF pleocytosis and FEP. The reason for the high prevalence (27%) of serum or CSF cell surface neuronal antibodies may be that our study targeted patients with FEP, and our subjects included many with anti-NMDAR encephalitis characterized by psychiatric symptoms.

As for the prognosis of encephalitis without FEP, a retrospective study focusing on the presence of seizures in different types of encephalitis showed that seizures have a significant relationship with 3-month outcomes [16]; however, the presence of psychosis was not mentioned. In our hospital cohort, an examination of 26 patients without FEP using a univariate logistic regression analysis revealed no significant prognostic factors (online suppl. Tables 2, 3). This suggests that there may be other different prognostic factors that could be used to differentiate the prognosis of encephalitis with and without FEP, and this will be the focus of further studies.

The present study used a nonspecific instrument to focus on a small subset of patients with FEP. In FEP patients, both larger areas, relevant to psychologists or generalists, and more specific brain areas, relevant to neurologists, are affected, and clinicians often wonder if FEP is derived from brain diseases such as encephalitis, especially in patients with anti-NMDAR encephalitis showing prolonged psychosis. In addition, many institutions cannot obtain early results for specific etiologies such as autoimmune antibodies, and many encephalitis cases have not been determined to have a known cause. Thus, the verification of the prognosis of overall encephalitis with FEP is vital.

Patients in our cohort received different types of steroid treatments; 11 patients received intravenous immunoglobulins or plasmapheresis (Table 1). This might have accounted for the differences in our prognostic results. Moreover, 9 patients had high fever (>38°C) at hospital admission, which might have influenced psychosis or decreased consciousness. Because of the retrospective nature of the study, we were not able to control for these variables. The additional univariate logistic regression analysis however indicated that both “the presence of other treatments” and “high fever >38°C at hospital admission” were not significantly associated with poor outcomes.

The association between brain regions and FEP has not yet been verified. In a study evaluating FEP subjects without encephalitis, combined functional MRI and single-voxel proton MR spectroscopy revealed altered functional connectivity strength and metabolic concentrations within the ventromedial prefrontal cortex [17]. Another 6-year follow-up study in FEP patients demonstrated bilateral hippocampal increase on MRI that occurred in a minority of patients following their FEP, which was associated with a good outcome [18]. However, the brain regions responsible for FEP in encephalitis remain uncertain. In anti-NMDAR encephalitis, characterized by prominent psychosis as a clinical presentation, the antibody reactivity is found in the entire brain and predominantly in regions with high neuronal and NMDAR density, such as the hippocampus [19]. A PET study reported temporomesial hyperintensities due to abnormal glucose metabolism in patients with mild to moderate and severe anti-NMDAR encephalitis [20]. Future studies such as prospective studies using neuroimaging techniques need to be conducted to confirm the brain region responsible for FAP in patients with encephalitis.

To the best of our knowledge, the present study is the first to identify a prognostic factor against overall subacute encephalitis with FEP. However, this study did have some limitations. Owing to the retrospective design of the study, 8 out of 22 patients without a confirmed etiology did not undergo neuronal antibody testing or PCR for HSV. Second, the final follow-up outcome was not analyzed, especially for anti-NMDAR encephalitis with long disease duration. Our use 2 months after the onset as a clinical outcome can provide clinicians with clues regarding the clinical course in the immediate future. In this study, prognosis of encephalitis was assessed using the modified Rankin scale. Prognosis of encephalitis can differ because of its various symptoms such as memory disorder, higher dysfunction, and motor paralysis. Currently, there is no prognosis index specific to encephalitis in consideration of these symptoms. However, the modified Rankin scale has been widely used to measure the prognosis of encephalitis [21]. Third, there is no comparison between the prognostic factors for FEP with encephalitis and FEP without encephalitis. Finally, this was a single-center study with a small sample size.

CSF examination, which is the most reliable method of detecting inflammation in the central nervous system, is not recommended for the routine diagnostic workup of patients with psychosis [1, 4, 5]. However, some classical psychotic manifestations respond to immunotherapy [5]. The necessity of CSF analysis depends on the individual case. For example, a patient with FEP with a neurological soft sign or feature should be considered for CSF examination [1]. The use of mechanical ventilation is a poor prognostic factor at 2 months after the onset of subacute encephalitis with FEP, and female sex may be a risk factor for unfavorable development of the disease. Nevertheless, clinicians should be aware that etiology is the best predictor of outcomes.

We are extremely grateful to all neurologists of Nara Medical University who did clinical practice for patients with encephalitis.

This study protocol was reviewed and approved by Medical Ethics Committee of Nara Medical University, approval No. 3126. The study has been granted an exemption from requiring written informed consent from Medical Ethics Committee of Nara Medical University, and the informed consent was given via opt-out. The appropriate participants’ informed consent in compliance with the Helsinki Declaration was obtained. No investigations or interventions were performed outside routine clinical care as this is without experimental intervention into routine care.

The authors report no conflicts of interest related with our paper.

There was no funding related with our paper.

Hiroshi Kataoka: substantial contributions to conception and design or analysis and interpretation of data. Hiroshi Kataoka, Hitoki Nanaura, Takao Kiriyama, and Kazuma Sugie: substantial contributions to drafting the article or revising it critically for important intellectual content. Hiroshi Kataoka, Hitoki Nanaura, Takao Kiriyama, and Kazuma Sugie: final approval of the version to be published.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.